Re: Oddness piping mkisofs into recent cdrecord versions

[scdbackup@gmx.net]

Hi,

Bill Davidsen:
> I can't replicate the "no TAO" with a CD,

This is the best outcome we can wish for.
A non-TAO CD burner would be very annoying.


> Question: what does "-dvd-compat" really do?

You mean in growisofs ? That's always interesting
to read :))

On the first hand it sets variables "poor_man" and
"dvd_compat".

poor_man is still a bit of a riddle to me, because
it is set to 1 at several occasions and its usage is 
manifold. It seems to apply to situations where a
real burner device is the target. All my use cases
with growisofs have (poor_man!=0).
Etymologcally "poor man" might stem from Andy's
2.4 kernel patch (= rich man, has block device suitable
to dd to media) and the "builtin-dd" for poorly
courageous people like me who prefer to run a
consumer kernel. From growisofs.c:
  * - "poor-man" support for those who don't want to recompile the
  *   kernel;


My (probably incomplete) idea of dvd_compat is that it
causes a DVD+RW to get formatted thoroughly ("de-iced"),
that it forces non-multisession for most other DVD+/-
media and then closes these.
It also does _something_ with Double Layer DVDs.

I refer to document mmc5r03c.pdf . Current is 
  http://www.t10.org/ftp/t10/drafts/mmc5/mmc5r04.pdf
It seems that MMC-6 has started recently.
Andy refers to MMC-4 in some of his program remarks.

dvd_compat:

- There is some some reservation against it in growisofs.c:
    /* never finalize disc at multi-sessioning DVD±R recordings...
    /* ... except when filling the media up:-)

- With DVD+RW it forces 04h FORMAT UNIT with format type 
  26h and restart bit. I.e. an eventually interrupted
  format from previous writing is resumed even if the
  burn job would fit into the already formatted area.
    MMC-5, 6.5.4.2.14

  It prevents the stop of de-icing via 5Bh CLOSE TRACK/
  SESSION, close function 000b "Quick Stop Background Format"
  at the end of a DVD+RW burn. De-icing makes addressable
  areas on the media which are yet unused.
    MMC-5, 3.1.17
    MMC-5, 6.3.3.6.2
  
- With DVD+R there is some deprecation of dvd_compat:
    * - 'growisofs -M /dev/cdrom=/dev/zero', this is basically a counter-
    [...]
    * - disable -dvd-compat with -M option and DVD+R, advice to fill up
    *   the media as above instead;

  If set, it disables multi-session in mode page 5
  and sets the write type to SAO.
    MMC-5, 7.5

  At the end of a DVD+R write it causes 5Bh CLOSE TRACK/
  SESSION, close function 110b "Finalize".
    MMC-5, 6.3.3.4.5
  rather than 010b "Close Session"
    MMC-5, 6.3.3.4.3

- With DVD+R DL : close function is 101b "Finalize with
  Minimal Radius"
    MMC-5, 6.3.3.5.5
  rather than 010b "Close Session"
    MMC-5, 6.3.3.5.3

  if (profile==0x2B && next_track==1 && dvd_compat && leadout)
     plus_r_dl_split((cmd,leadout);
  which issues a command BFh SEND DISC STRUCTURE.
    MMC-5 6.43 (i did not read this yet)

- With sequential DVD-R* and multiple -dvd-compat, variable
  "is_dao" gets set to 1. It is set automatically if feature
  0021h "Incremental Streaming" is not offered by the drive.
  (MMC-5: "On CD media, this is known as packet recording")
  is_dao causes write type 02h "SAO"
    MMC-5, 7.5.4.9
  and other behavioral modifications, like suppressing
  5Bh CLOSE TRACK/SESSION with close function 001b 
  "Close a Logical Track".
    MMC-5, 6.3.3.3.2
  (I did not get to sequential DVD recording yet. Still
   exploring DVD-RW Restricted Overwrite where i try to
   make an own theory how to manage sessions. Somehow i
   do not recognize the growisofs doings in the specs -
   although growisofs works well and my own derived sketch
   seems to work too ...)

  It disables multi-session with DVD-R DL.
    MMC-5, 7.5

- Various text messages talk of closing "disc" rather
  than closing "session".


Hey Andy: How good is my score ?
How much did i miss and what did i misunderstand ?


Have a nice day :)

Thomas